mirror of
https://github.com/ggml-org/llama.cpp.git
synced 2026-10-02 02:47:26 -05:00
This lets the copy to the destination device use the host-visible vidmem optimization.
This lets the copy to the destination device use the host-visible vidmem optimization.